Physical properties of celery juice

Transcription

Physical properties of celery juice
Physical properties of celery juice
A.K. LAU1, A.C. MARCH1, K.V. LO1 and D.B. CUMMING2
department of Bio-Resource Engineering, University of British Columbia, Vancouver, BC, Canada V6T 1W5; and Food
Processing Section, Agriculture Canada Research Station, Summerland, BC, Canada VOH1Z0. Contribution No. 745. Received
27 November 1989; accepted 11 June 1991.
Lau, A.K., March, A.C, Lo, K.V. and Cumming, D.B. 1992. Physical
ployed as a first stage process with other technologies like
properties of celery juice. Can. Agric. Eng. 34:105-110. Before
undertaking a project to design a pilot-scale freeze concentration
system to dewater celery juice, laboratorystudies were carried out to
freeze concentration completing the concentration system. As
a result of significant advances made in improving the freeze
concentration process (Deshpande et al. 1984; Smith and
Schwartzberg 1985; van Pelt and Swinkels 1986), freeze con
centration is now readily adopted in the food industry for
determine the physical, thermal and rheological properties of celery
juice at various total solids concentration and various temperatures.
The properties measuredwere solids content, specificgravity, electri
cal conductivity, freezing point and melting characteristics, specific
heat, thermal conductivity and dynamic viscosity. Regression analyzes
were carried out for correlating specific gravity, electrical conductivity
and specific heat with total solids content. Freezing point depressions
of celery juice were compared to similar materials. Values of thermal
conductivity were reported for a range of solids concentrations. Rheo
logical properties of celery juice were represented by a family of
dynamic viscosity curves.
Avant la conception d'un systeme pilote de concentration par
congelation, destine a deshydrater le jus de celeri, des etudes de
laboratoire ont ete effectuees pour determiner les proprietes physiques,
thermiques et rheologiques du jus en fonction de diverses concentra
tions de matieres solides et de temperatures. Les proprietes mesurees
etaient les suivantes: teneur en matieres solides, densite, conductivite
electrique, point de congelation, caracteristiques de fusion, chaleur
massique, conductivite thermique et viscosite dynamique. Des analy
ses de regression ont egalement ete effectuees pour correler la densite,
la conductivite electrique et la chaleur massique avec la teneur en
matieres solides. Puis on a compare les abaissements du point de
congelation du jus de celeri a d'autres matieres similaires. Des valeurs
de conductivite thermique ont ete determinees pour toute une serie de
concentrations de matieres solides, et les proprietes rheologiques du
jus ont ete representees par des courbes de viscositedynamique.
INTRODUCTION
Large quantities of celery are harvested from mid-July to
mid-November in the Fraser Valley area of British Columbia,
Canada. The celery harvested is used only to supply the fresh
produce market, and up to 50 percent of the crop is either left
in the field or designated for post-harvest processingrejection.
Celery juice can be extracted from this otherwise wasted ma
terial to be used as a component in commercial vegetable
juices and soups, thus providing a further source of revenue. A
frozen celery juice concentrate is needed to reduce the bulk to be
transported to distant markets, for example, Ontario where the
commercial juices and soups are blended and manufactured.
Three main groups of technically feasible processes used
for the concentration of liquid food are evaporation, reverse
osmosis andfreeze concentration. Forjuicescontaining aroma
volatiles, the concentration process has to be highly selective.
Where the high osmotic pressure of fruit juices precludes their
concentration to the required level of total solids, Medina and
Garcia (1988) recommended that reverse osmosis be em
CANADIAN AGRICULTURAL ENGINEERING
manufacturing high quality products.
As part of a project funded under the Energy Research and
Development in Agricultureand Food program,the feasibility
of freeze concentrating celery juice was studied. To adapt and
test at pilot scale a freeze concentration system for dewatering
celery juice, laboratory studies were required to delineate the
physical properties of celery juice. In particular, for predicting
freezing rate, crystal growth rate and refrigeration capacity,
data on thermal properties and rheological properties were
needed.
The dewatering capacity of a freeze concentration system is
highly dependent upon the initial and final dissolved solids
content which affect the liquid viscosity at the freezing point
temperature. Solute concentration is also a major factor that
affects the nucleation rate and growth of ice crystals in a mixed
suspension. The specific gravity and thus the bulk density of
celery juice is needed for the calculation of the pumping power
requirements, as well as for the design of the ice/concentrate
separation device in a freeze concentration system. Celery
juice is rich in salts (Butfkus 1978) and is a good conductor.
The electrical conductivity of celery juice determined as a
function of dissolved solids concentration can provide a sim
ple and effective means for monitoring the product
concentration continuously. An output signal from the meter
could be used in a control circuit to maintain product unifor
mity by controlling internal flow rates and heat-removal rates.
The freezing point (FP), specific heat and thermal conductivity
of solutions are needed parameters for the design of freezing
and thawing processes. The critical diameterof a crystal varies
direcUy with the FP of the solution duringcrystallization, andthe
FP of the concentrate has a significant impact on the power
consumption of the process. Viscosity valuesaid in the sizing of
pumps and aid in determining the applicability of a flooded wash
columnfor the separation of ice from the concentrate/ice slurry.
The purpose of this paper is to report the results of tests
undertaken to determine the following properties for celery
juice: solids content, specific gravity, electrical conductivity,
freezing point and melting characteristics, specific heat, ther
mal conductivity and dynamic viscosity. As juice
concentration increases progressively during the freeze con
centration process, these physical properties were evaluatedat
various total solids concentrations.
Vol. 34, No. 1, JANUARY/FEBRUARY/MARCH 1992
105
MATERIALS AND METHODS
freezing point for each level of concentration. The apparatus
Methods adopted in this study for measuring physical charac
teristics, thermal properties and rheological properties of
was precooled to one of five temperatures between -10°C and
liquid food have been thoroughly reviewed by Mohsenin
(1980,1986).
Initially, thejuicewas extracted byscrew press from celery
stalks which were rejected from the fresh market. Clarified
juice was produced by filtration with a Millipore 0.22 |nm
mean pore diameter filter. Juice was frozen and held at -20°C.
One barrel each of unclarified and clarified celery juice was
allowed to thaw completely and then was thoroughly mixed
prior to and during sampling. The samples were storedat 3°C
before testing.
For the purpose of substrate characterization, it was as
sumedthat freeze-drying would provide samples that were of
acceptably close quality to freeze-concentrated juice. Ten li
tres of single-strength juice was freeze-dried to a light brown
powder. Thepowder form of thejuice wasbiologically stable
and allowed for easy sample preparation for lab analysis. A
range of concentrations was prepared by adding powder in
measured quantities to 100 ml volumetric flasks and filling to
the line with room temperature distilled water.
All tests for physical properties used rehydrated juice sam
ples unless indicated otherwise, and duplicate or triplicate
sampleswere used in each test run for each physical property.
Solids content
The determination of total solids (TS) concentration in mass
percentage was carried out by drying a known mass of sample
inanoven at 105°C for24hours. These drysamples were then
placed in a muffle furnace at 550°C for 20 minutes to deter
mine the percentage of volatile solids (VS). Suspended solids
(SS) and volatile suspended solids (VSS) were determined
using the same drying procedure outlined above from the
suspended material filtered out of the samples with Whatman
glass microfibre 934-AH filters in a Buchner funnel vacuum
filtration assembly. The filters were conditioned in the furnace
at 600°C for 15 minutes to eliminate any trace moisture or
volatiles.
-18°C. Samples were tested according to the following proce
dure: Carbon dioxide cartridges were rinsed with distilled
water.Each cartridge, with an interior volume of about 10 ml,
was filled with 7.00 ml of a given juice sample having a
dissolved solids concentration between 5% and 30%. Since the
suspended solids concentration was low, the dissolved solids
(DS) and TS concentrations were essentially the same. The
cartridges were thenplaceduprightin a circularplastic holder
and the thermocouple leads from all samples were attached to
a datalogger (Digitek Model 1268). A Kaye model K140-4 Ice
Point Reference was used as a temperature reference. The
datalogger was activated as the samples were immersed in the
bath. Readings were recorded at 15 s intervals until the tem
perature of the samples approached that of the bath,
terminating the run. Runs were made both with and without
agitation of the samples.
A differential scanning calorimeter DSC (Perkin Elmer
ModelDSC-2) along with a thermal analysis data station were
used to further examine the melting characteristics of the
prepared juice samples.
Melting curves were obtained to gain some insight into the
phase change characteristics of the juice at various levels of
concentration. After each test, the graphical results were ana
lyzed to determine the melting point onset temperature, the
specific energy gain and the temperature corresponding to the
maximum melting rate for the sample.
Specific heat
The DSC also provided a rapid and dynamic method to deter
mine the specific heat of celery juice, as described by Otten et
al. (1980). The sample and the sapphire standard were forced
to simultaneously scan through a temperature range of -40 to
20°C at a constant rate of 10°C/min. The difference in the
power input between the two cells was a direct function of the
mass and specific heat of the sample. A specific heat value was
obtained from the DSC thermograms by dividing the heat
input rate in MJ/s by the temperature scanning rate in °C/s
and the sample mass in mg to yield the appropriate units of
Initially, both clarified and unclarified juice samples from
the original celery juice that had been frozen were used in the
above-mentioned test procedure. Subsequent measurements of
this parameter, during the evaluation of physical properties as
a function of total solids content only, made use of rehydrated
celery juice.
kJ-kg" •°C" . A comparison of the sample's thermogram (rate
Specific gravity
The thermal conductivity at the freezing point for various
concentrations of celery juice was calculated from the follow
ing empirical equation for fruit juices and sugar solutions
(Heldman and Singh 1981). This equation has incorporated the
effect of the thermal conductivity of water and of the solids
A 25 ml flask was weighed before and after filling to the line,
in turn, with room temperature (22°C) distilled water and
celery juice. A Mettler H6 digital analytical balance was used
to determine all masses to the nearest 0.1 mg. Tests were
performed in ascending order of concentration.
Electrical conductivity
Electrical conductivity (EC) was determined by use of an EC
meter (YSI Model 31) and probe. Unfiltered juice samples
prepared from freeze-dried celery juice powder were tested at
26°C.
Freezing point and melting characteristics
A refrigerated circulating bath was used to determine the
106
of energy input versus temperature) with that of the sapphire
standard of known specific heat allowed calculation of the
sample's specific heat.
Thermal conductivity
component in the product.
k = [326.575 + 1.0412-F/> - 0.00337-CF/T] [0.796
+ 0.009346-(100-75)]-10'3
(1)
where:
k =thermal conductivity (Wnn'^C1),
FP = freezing point (°C), and
TS = total solids (%).
LAU, MARCH, LO, and CUMMING
Once the specific heat and thermal conductivity are known,
thermaldiffusivity can be readily calculated.
Viscosity
A Cannon-Fenske Routine type viscometer was used to esti
mate theviscosity of celery juiceovera range of temperatures
anddissolved solidsconcentrations. Experiments werecarried
below 18%, and show a nearly linear response for EC vs TS.
The straight line portion may be mathematically described by
EC = 10.377 +2.300 TSC
(R2 =0.982)
(5)
where EC = electrical conductivity (mmho).
out at temperatures down to the freezing point or its vicinity.
The temperature control jacket was connected to a Lauda
RCS-6D refrigerated circulating bath for very stabletempera
turecontrol (±0.02°C). A #100Z13 tube witha recommended
range of viscosities from 3 to 15-10 Pa-s was used for all
runs. The efflux time (s) was measured to the nearest 0.1 s and
the kinematic viscosity value (m /s) was calculated by multi
plying the efflux time by a temperature-compensated tube
constant. The kinematic viscosity can easily be converted to
dynamic viscosity (Pa-s) via the specific gravity.
RESULTS AND DISCUSSION
Solids content
The solids analysis for the unconcentrated juices is presented
in Table I. The unclarified juice had a TS concentration of
3.5%comparedwith 2.6% for the clarified juice; the filtration
of the whole, unconcentrated juice reduced the TS by 24% and
the VS by 29% by removing 98% of the SS. The mass percent
age solids in the clarified and unclarifiedjuices were related
according to the regression equation:
(JT = 0.966)
TSC = 0.409 + 0.806 TSU
(2)
where:
TOTAL SOLIDS (%)
Fig. 1, Electrical conductivity ata temperature of26°C.
Since 18% TS represents a concentration factor of seven,
while the most likely final concentration for the juice in a
freeze concentration process is four (10.5% TS), this appears
to be a very useful method for determining solids concentra
tion and providing control loop feedback signals during
operation. Temperature compensated probes allow readings
taken at temperature from as low as -5°C to as high as 100°C
TSC = total solids in clarified juice (%), and
TSU = total solids in unclarified juice (%).
to be converted to standard readings at 26°C.
Total
Volatile
Suspended
Volatile
solids
solids
solids
suspended
Freezing point and melting characteristics
Based on the transient temperature measurements, the freezing
point (FP) temperatures were taken to be the highest ones
achieved, as obtained from the crystallization exotherms fol
lowing supercooling. The degree of supercooling generally
solids
decreased as the TS concentration increased and as the FP of
Table I. Results of solids analysis for unconcentrated
celery juice
(%)
(%)
(%)
(%)
Clarified
2.62±0.07
1.70±0.04
0.0098±0.0005
trace amt
Unclarified
3.47+0.03
2.39±0.05
0.44+0.01
0.42+0.01
Specific gravity
Experimental results of specific gravity were analyzed by
means of a linear regression of specific gravity on total solids
content, which returned the equations:
SGC= 1.000 + 0.006 TSC
(/T = 0.984)
(3)
SGU = 0.997 + 0.005 TSU
(/T = 0.971)
(4)
where:
SGC = specific gravity of clarified juice, and
SGU = specific gravity of unclarified juice.
Electrical conductivity
The electrical conductivity readings, as demonstrated in Fig. 1,
are quite sensitive to changes in the level of total solids content
CANADIAN AGRICULTURAL ENGINEERING
the solutions approached the bath temperature. Crystallization
exotherms which were rounded in shape with no isothermal
portion, may have been peaking at some temperature below
the actual FP of the solution. Therefore, the refrigerated circu
lating bath method yielded at best a close approximation of the
actual FP, which could only be determined precisely by allow
ing steady-state conditions to prevail after each temperature
increment or decrement. The latter procedure is exceedingly
time-consuming, since at each operation the proportions of ice
and solution must approach equilibrium. However, the FP
temperatures obtained from these tests are adequate for the
design and start-up of a freeze concentration system having a
TS monitor and control loop with which to fine-tune the bulk
temperature to produce the desired concentration.
A typical temperature-time curve for 25% TS celery juice is
shown in Fig. 2. Following the onset of crystallization at
-14°C, the releasedheat of crystallization causedthe tempera
ture to rise to the FP of -8°C. Figure 3 depicts the FP
temperatures at various TS levels. For comparison purpose,
the FP curves for coffee extract, wine and blackcurrant juice
(Thijssen 1974) are shown in the same graph. The rapid low-
Vol. 34, No. 1, JANUARY/FEBRUARY/MARCH 1992
107
CELERY JUICE 10%
MASS: 8.37 mg
SCANRATE: 10.00 dog/mln
PEAK FROM: -34.0 J °C TO 7.17 °C
MAX: -0.96 °C
ONSET: -7.28 °C
kJ/kg: 268
TEMPERATURE ("C)
TIME
(min)
Fig. 2. Typical freezing curve of celery juice at 25%
total solids.
CELERY JUICE 25%
MASS: 6.72 mg
SCAN RATE: 10.00 dog/mln
PEAK FROM: -34.78 <C TO 2.67 X
ONSET: -11.98 °C
kJ/kg: 214
I
TEMPERATURE (°C)
cdwrJufc*
10
IS
20
CELERY JUICE 40%
TOTAL S0UDS (%)
MASS: 5.95 mg SCAN RATE: 10.00 dog/mln J
Fig. 3. Freezing point depression as a function of total
solids content (celery juice: present study; other
materials: Thijssen 1974).
PEAK FROM: -36.28 °C TO -2.21 °C
ONSET: -18.88 "C
kJ/kg: 154
ering of the FP for wine is due mainly to the increasing
concentration of alcohol.
Figures 4(a), (b) and (c) illustrate the melting curves and
contrast the maximum melting points, peak shapes and areas
for 10%, 25% and 40% TS, respectively. The shape of the
peaks shown are very similar, while the peaks are shifted to the
left for increasing solids content, demonstrating the FP depres
sion due to the higher TS concentration. The melting peak is
seen to cover a wide temperature range rather than occurring
at a fixed temperature as is the case for an operating freeze
concentration system. The differential scanning calorimeter
results are based, however, not on the melting of pure ice
crystals in various TS concentrations, but on the melting of a
mixture of ice crystals and solution.
Some irregular patterns are observed in Figs. 4(a) and (b)
-40
-30
-20
-10
0
10
20
TEMPERATURE (°C)
Fig. 4. Differential scanning calorimeter (DSC-2) melting
curves (a) solids content = 10% (b) solids
content = 25% (c) solids content = 40%.
when temperature is at approximately -12°C. The slight flat
crease, which can be explained by assuming the occurrence of
a 'glass transition phenomenon' at this temperature. The sam
tening of the slope indicates that a smaller amount of heat was
required to be added for a constant rate of temperature in-
calorimetry run, thus the solution would freeze before any
108
ples were quickly cooled to -40°C at 360°C/min prior to each
LAU, MARCH, LO, and CUMMING
CO
9
20
-J
o
CO
-J
ss
o
0 ]
-20
i
-18
1
-16
1
-14
I
-12
-10
-8
-6
40
50
60
TOTAL SOUDS (%)
TEMPERATURE (C)
Fig. 6. Mean specific heat of celery juice as a function of
Fig. 5. Melting characteristics as derived from the
total solids.
melting curves.
significant crystal growthcould take place. Upon warming, as
the temperature reaches -12°C, the solution has melted to the
point where free water molecules have enough mobility to
change from a random, 'glassy' state to a lower energy crys
talline state. The heat of crystallization thus released
momentarily reduces the rate of heat input. It shall be noted
that this postulated glass transition is not likely a factor of
concern in a freeze concentration process, since the FP is
always approached from the high side.
The melting point onsets and the temperature corresponding
to the maximum melting rate of the sample as derived from the
melting curves are illustrated in Fig. 5. That the best-fit lines
for the two variables diverge as percent TS increases, indicates
the lowering of the FP and the simultaneous broadening of the
melting peak with increasing percent TS. This phenomenon
explains the difficulty encountered in establishing the FP of an
increasingly impure solvent by the exotherm-plateau method
as previously discussed.
Specific heat
the measured freezing point at each solids content.
The results of specific heat and thermal conductivity dem
onstrated that their values were not significantlydifferentfrom
those of water, for total solids content of celery juice up to 18%
(7-fold concentration factor).
Viscosity
Results from the Cannon-Fenske Routine Viscometer are pre
sented in Fig. 7 in the form of a family of dynamic (absolute)
viscosity curves for celery juice. These results augmented
Kuprianoff's (1966) results for fruit juices. For the range of
temperatures investigated, at 10.5% and 18% TS levels (that
is, 4-fold and 7-fold concentration factors), the viscosity of
celery juice is about 40% and 85%, respectively, greater than
that ofwater (1.8-10 3Pa-s at 0°C).
Table II. Thermal conductivity of celery juice at its
freezing point
content
Freezing
point
conductivity
(%)
(°C)
QNnaloCl)
0
0.0±0.05
0.565
kJ-kg'1-^"1 was obtained for water, which compares very
well with the reported value of4.18 kJ-kg"1-^"1 (Kreith and
5
-1.0±0.24
0.549
10
-2.2±0.26
0.531
Black 1980). The following linear equation was fitted to the
experimental data:
15
-3.6±0.33
0.513
20
-5.710.68
0.495
25
-8.210.97
0.476
30
-9.111.80
0.459
Solids
Specific heat of celery juice is plotted in Fig. 6 as a function of
percent TS. Mean specific heat values for each TS content
were calculated from test results with temperature ranging
from -40°C to 20°C.The experimental runs included measure
ment taken at 0% TS and an average value of 4.21±0.17
cp = 4.22- 0.03 TSC
(Rz = 0.958)
(6)
where cp = specific heat (kJ-kg" -°C" ).
CONCLUSIONS
Thermal conductivity
Table II shows the thermal conductivity for a range of TS
content of celery juice at the corresponding temperature of its
freezing point as interpolated from Fig. 3.
As TS changes from 0% to 18%, thermal conductivity only
decreases by 11% from 0.565 to0.502 W-m^C1. The ther
mal conductivity was not sensitive to the standard deviation of
CANADIAN AGRICULTURAL ENGINEERING
Thermal
Tests on physical characteristics, thermal properties and rheo
logical properties were carried out for a range of total solids
concentration and at various temperatures. Specific gravity,
electrical conductivity and viscosity were found to increase
with an increase in solids content of the juice samples, while
freezing point temperature and specific heat vary adversely
with solids content. Linear equations were fitted to experimen-
Vol. 34, No. 1, JANUARY/FEBRUARY/MARCH 1992
109
g
\v
%TS
O 0.0
•
4.2
D 7.9
•
V ^
V**^
o^>
Salunkhe. 1984. Freeze concentration of fruit juices.
Critical Reviews in Food Science and Nutrition
20(3): 173-248.
11.6
A 15.5
Heldman, D.R. and R.P. Singh. 1981. Food Process
O 19.2
Engineering, 2nd ed. Westport, CT: AVI Publishing
V
v^
Deshpande, S.S., M. Cheryan, K.S. Shridhar and D.K.
25^
X Ff?«dngJ,olnts
Company Inc.
Kreith, F. and W.Z. Black. 1980. BasicHeat Transfer, 2nded.
V^s
New York, NY: Harper and Rows.
^•ET"
^N\ 7
Kuprianoff, J.M. 1966. Food properties point way to quality.
Food Engineering 2:68-69.
Medina, B.G. andA.Garcia III. 1988. Concentration oforange
juice by reverse osmosis. Journal of Food Process
Engineering 10:217-230.
TEMPERATURE (*C)
Fig. 7. Dynamic viscosity vs. temperature.
taldata, forcorrelating specific gravity, electrical conductivity
and specific heat with total solids content.
Although celery juice hada largerfreezing pointdepression
compared to coffee extract and blackcurrant juice, its low
solids concentration and viscosity still renders it suitable for
the freeze concentration process. A maximum viscosity of
200-10" Pa-s and total solids concentration of 45-55% for the
concentrated product had been recommended by Thijssen and
van Oyen (1977). Increased viscosity can make pumping the
concentrate and washing the ice very difficult. The low total
solids content of thejuice also implied that the thermal prop
erties can be assumed to be the same as for water for design
purposes.
ACKNOWLEDGEMENTS
Mohsenin, N.N. 1980. Thermal Properties of Food and
Agricultural Materials, New York, NY: Gordon and
Breach Science Publishers.
Mohsenin, N.N. 1986. Physical Properties of Plant and
Animal Materials, 2nd ed. New York, NY: Gordon and
Breach Science Publishers.
Otten, L., G.Y.I. Samaan and G.O.I. Ezeike. 1980.
Determination of the specific heat of agricultural
materials: Part I. Continuous adiabatic calorimeter.
CanadianAgricultural Engineering 21:65-71.
Smith, C.E. and H.G. Schwartzberg. 1985. Ice crystal size
changes during ripening in freeze concentration.
Biotechnology Progress 1(2):111-120.
Thijssen, H.A.C. 1974. Freeze concentration. In Advances in
Preconcentration and Dehydration of Foods, ed. A.
Spicer, 115-149. New York, NY: John Wiley and Sons.
Thijssen, H.A.C. and N.S.M. van Oyen. 1977. Analysis and
economic evaluation of concentration alternatives for
The authors thank Neil Jackson for his technical assistance,
and East Chilliwack Agricultural Co-op for their coopera
tion in supplying the juice samples. Funding provided by
liquid foods - quality aspects and costs of concentration.
Journal ofFood Process Engineering 1:215-240.
Agriculture Canada through ERDAF contract no.
van Pelt, W.H.J.M. and W.J. Swinkels. 1986. Recent
04SB.01916-3-EP19 is also gratefully acknowledged.
REFERENCES
Buttkus, H.A. 1978. Celery leaf juice: Evaluation and
utilization of a product from harvest debris. Agricultural
and Food Chemistry 26(4):827-837.
no
developments in freeze concentration. In Food
Engineering and Process Applications, Vol. 2, Unit
Operations, ed. M. Le Maguer and P. Jelen, 261-273.
London, England: Elsevier Applied Science Publishers.
LAU, MARCH, LO, and CUMMING